Attribution has always been messy. But three forces are making it worse at once: cookie restrictions, AI search, and zero-click behavior.
Ted Coxworth has a front-row seat to all of it. As President of Extensor, a marketing measurement firm, he works with brands to assess channels ranging from Google and Meta to billboards and radio. His tool of choice for cutting through the noise: media mix modeling.
The Attribution Problem Just Got Harder
On the cookie side, privacy regulations have tightened for the better part of a decade, and Apple led the push to let users opt out of tracking entirely. The result? Direct and organic traffic in multi-touch attribution systems has ballooned.
"That used to be 10%, and now it's 40%," Ted said. "That doesn't quite make sense."
That gap is why multi-touch attribution has become increasingly unreliable. And it's why more brands are turning to an older tool: media mix modeling.
What MMM Actually Is
Media mix modeling, sometimes called marketing mix modeling, measures the effectiveness of a marketing program by looking at correlations across time periods and markets.
"So maybe you're spending a lot on Google in Salt Lake City, and maybe you're spending a little bit less on Google in the Houston DMA," Ted explained. "What do the sales look like in those two places across time? And can we see a correlation? When you spend more on Google, you get more sales."
The key difference from MTA: MMM does not track individuals. It looks at patterns of spend, impressions, and sales across channels and markets, then measures how they relate to the KPI that matters, whether that's revenue, leads, or something else.
One major strength is that non-marketing factors can live in the model too. Pricing, competitive behavior, macroeconomics, even weather. Ted pointed to bug spray as an early example from his Nielsen days: pull in weather data and sales behavior suddenly makes sense. The same dynamic shows up in the legal industry, and in tourism. In a project he worked on with Paxton Gray, pulling in visa application data had a real impact on the model's accuracy.
Ted also noted a counterintuitive finding on competitive data: "The overwhelming majority of cases show that more competitive spending often results in lift for your brand." More category awareness, it turns out, tends to raise all boats.
MMM is not a new idea. Ted has been building these models since 2013, and the technique itself has been around for roughly 70 years.
Enterprise AI Discoverability Series
Buyers are discovering, evaluating, and comparing brands through AI. 97th Floor CEO, Paxton Gray, is joining WordPress VIP CMO, Jodi Cerretani, for a three-part series on what that means for enterprise brands.
Three Tools, Three Jobs
Ted Coxworth frames MMM, MTA, and incrementality testing as a three-part system, not competing options.
MMM is the big-picture tool. It shows how everything connects, accounts for non-marketing factors, and is useful for media planning and scenario analysis. The limitation is that it is based on correlation, not causation.
Incrementality testing closes that gap. Run a test with matched geographic markets, spend on a new channel in one and not the other, and if sales go up, you have a causal link. "You can say pretty clearly, hey, that additional spend drove those sales," Ted said. Tests typically take four to six weeks.
MTA handles day-to-day operations. It is always on and gives fast results. "If you can say this channel gets 10 last-touch sales, but I know it probably drove 20 total sales, you can look at your MTA, look at that last-touch number, and extrapolate out and say, is my holistic CPA for this channel where it needs to be?"
The workflow Ted recommends: use MMM to find opportunities, use incrementality tests to prove them out, and use MTA to manage execution.
First-Party Data Is Not Enough on Its Own
Many brands are leaning into first-party data as the answer to tracking loss. Coxworth sees value in it but flags a consistent limitation, especially with customer surveys.
"Typically, those are going to privilege certain marketing channels," he said. Surveys tend to act like a last-touch model. And most do not let a customer name more than one channel. "They'll say Google, or they'll say the billboard on I-15, and that's it." But in most purchase paths, more than one thing drove the sale.
"I don't think any of them do a great job of telling the holistic story that MMM can tell."
AI Search: Even Harder to Measure
If cookie loss made attribution messy, AI search is making it messier.
In traditional SEO, a brand could track impressions, rankings, keyword volume, and click behavior. In AI overviews, the data available is far thinner. "The closest we've been able to get is to say, for a set of prompts, on a certain date, were you in the citation list or not?" Ted said. "Which is a little bit different than saying, this many sets of eyes saw your brand."
That gap is exactly where MMM becomes useful. Because you cannot link an individual purchase to an AI overview, looking at correlations between citation data and sales becomes the best available tool.
Ted has already seen this play out in client work. For some brands, a drop in organic traffic is offset by a rise in AI overview citations, and the MMM catches it. For others, the picture is murkier. "I think still more to come when it comes to measuring AIO."
He also addressed what happens when major platform shifts disrupt the model entirely. If organic search loses its impact because AI overviews replace the ten blue links, that shows up in the model fast. "That relative importance is just going to start coming down very crisply, run after run after run." Options range from de-prioritizing the channel to removing it from the model altogether.
The Lower-Funnel Trap
Paxton raised a McKinsey study showing that paid media as a share of total marketing spend climbed from roughly 20% in 2023 to 30% in 2025, the fastest-growing category. Ted did not find that surprising.
"Once you get on that treadmill, it's very hard to get off," he said. "Because you're not really driving a lot of demand, you're just capturing it, which can be very efficient at first. But once that demand starts to slow down, all of a sudden it's not so efficient anymore."
The issue is that platforms like Google and Meta have sophisticated targeting tools for people about to buy. They will tell you an ad was served and a conversion happened. What they will not tell you is whether that sale was incremental, or whether the customer would have found you anyway.
"Investing more in harvesting demand is not the same as investing more in generating demand."
MMM and incrementality testing both help brands see what mid- and upper-funnel investment is actually doing for the KPI that matters, and make the case for spending there.
How to Get Started
For marketing leaders who want to move in this direction, Ted laid out a practical sequence.
First, know your own sales data. Being able to pull clean sales numbers by day and by geography is the foundation.
Second, make sure tracking across the marketing program is solid. Impressions and spend need to be relatable.
Third, run a test. In-platform tests from Meta and Google are accessible starting points. A geo holdout experiment is more rigorous. Either way, starting with a test builds organizational buy-in and produces results you can act on fast. "With a test, you can get really solid actionable insights very, very quickly and use that to build some momentum."
MMM itself is a bigger investment. Ted said it typically takes four weeks or more to produce results.
The One Thing to Change
When Paxton asked what one shift would make brand-side marketers more effective, Coxworth was direct.
"Figure out a way to show that your marketing efforts require more than one touch or more than one platform to lead to a sale."
Most marketers know that is true. The challenge is proving it to finance and executive leadership. Once it is proven, the argument for investing across the funnel becomes a lot easier to make.
Resources:
Learn about Ted's work at tapindigital.com
Connect with Ted on LinkedIn: www.linkedin.com/in/ted-coxworth-4330568b
Reach Ted directly: ted@tapindigital.com
Connect with Paxton on LinkedIn: https://www.linkedin.com/in/paxtongray/
Looking for an agency that'll be worth the investment? 97th Floor creates custom, audience-first campaigns that drive pipeline and conversions. Get started here: https://97staging.com/lets-talk/.
About Ted Coxworth:
Ted Coxworth is President of Extensor, a marketing measurement firm that shows brands what's working, what's not, and how to get more from their media budget. He works with both in-house marketing teams and agencies to assess channels ranging from Google and Meta to billboards and radio, giving him a front-row seat to the measurement challenges across B2B and B2C businesses alike. He believes every marketer deserves an honest picture of how their budget works.
